Phonology of Telugu

HANDBOOK OF PHONOLOGICAL DATA
FROM A SAMPLE OF THE WORLD'S LANGUAGES

A Report of the Stanford Phonology Archive

Compiled and edited by

John H. Crothers
James P. Lorentz
Donald A. Sherman
Marilyn M. Vihman

1979 by The Board, of Trustees of the Leland Stanford Junior University

All rights reserved
Printed in the United States of America

PAGE 232 STANFORD PHONOLOGY ARCHIVE

VOLUME 1 — SEGMENT INVENTORIES, GENERAL COMMENTS, FOOTNOTES

Telugu

415

415 Telusu

415 Telusu

415 01 p

415 02 p-iong

415 03 p-aspi rated 30
( loan)

415 04 b

415 05 b-long

415 06 b-breathy voice 30
( loan)

415 07 t 01

415 08 t-long

415 10 d 01

Eeth] 60

415 11 d-long

415 12 d-breathy voice 30 31
( loan)

415 13 t-retroflex 02

415 14 t-retroflex-long

415 15 t-retroflex-aspi rated 30
( loan)

415 16 d-retroflex 02

tr-flap-retroflex] 03 60

415 17 d-retrof lex-long

415 18 d-retroflex-breathy voice 30
( loan)

415 19 k

Ik-lax] 60

415 20 k-long

415 21 k-aspi rated 30
( loan)

415 22 g

rg-lax) 60

415 23 g-lcng

415 24 g-breathy voice 30
( loan)

415 25 t/s 32

*tt/s-hacekJ

415 26 t/s-long

Ct/s-hacek-long) 61

415 27 d/z

td/z-hacek] 61

415 28 d/z-long

415 Telugu

Id/z-hacek-long] 61

29 t/s-hacek 32 61
(alio, loan)
*/t/s/

33 f 33
( loan)

35 s 3 *

( restricted)
*[s-hacekl

37 s-hacek-retrof lex

39 s-hacek 31 *

(tag(-), free, alio)
*/s/

41 m
Iw-nasalizedJ 62

42 m-long

43 n 06 35 3S
[eng) 63
tn-palatalJ 61 *

44 n-long

45 n-retrof lex 35

47 1
ll-dental) 6 *

48 1-long

49 1-retroflex

50 1-retrof lex-long

51 r-flap 07

52 r-trill-long

53 h-voice

51 i
*Uota)

52 iota 09 66
(alio, surface)

*/i/

53 i-long

*[ iota-long)

54 iota-long 66
(alio, surface)

*/i-long/

55 e
*[epsi Ion)

56 epsilon 10 66
(alio, surface)

*/e/

57 e-long
*Eepsi Ion- long]
*[ ash-long]

58 epsi lon-long 66
( alio, surf ace)

*/e-long/

59 ash-long 11 12 37
(tag(-),free)

*/e-long/

60 schwa

* i a 1 pha -unrounded]
lash-dot] 13 67

61 alpha-unrounded 1 ** 66
(allo,surface)

*/schwa/

62 a-long
*talpha-long]

ta-front-long] 15 68

63 alpha-long 66
(alio, surface)

*/a-long/

64 u 16 17
*tupsilon)

65 upsilon 18 66
(alio, surface)

*/u/

66 u-long 17
*tupsi lon-long]

67 upsi lon-long 66
(alio, surface)

*/u-long/

68 o 17
*to-mid]

69 o-mid 66
(alio, surface)

*/o/

70 o-long 17
*to-open-long]

71 o-open-long 66
(alio, surface)

*/o-long/

72 yod

73 beta-approximant 19
tw] 69

74 beta-approximant-long 20

PAGE 233 STANFORD PHONOLOGY ARCHIVE

VOLUME 1 SEGMENT INVENTORIES, GENERAL COMMENTS * FOOTNOTES Telugu

415

415 $a Telugu $d Dravidian $e SE India (Andhra Pradesh) $f 50 million $g Merrltt Ruhlen$g Gary

Holland (review) $g John Crothers (editor)

415 $a Lisker, Leigh $b 1963 $c Introduction to Spoken Telugu $g New York* ACLS

415 $a Krishnamurti, Bhadriraju $b 1961 $c Telugu Verbal Bases $f (University of California

Publications in Linguistics* Vol.24) $g Berkeley and Los Angeles: University of California
Press

415 $a Kelley, Gerald $b 1963 $c Vowel phonemes and external vocalic sandhi in Telugu. $d Journal

of the American Oriental Society* 83.67-73

415 $a ACCENT $A Neither Lisker nor Krishnamurti mentions word accent patterns. Kelley (p.67f ) has

a system of indicating "primary" and "tertiary" stress which seems in general to put stress on.
the first syllable of a word or phrase* while allowing for some phrases to begin with
"tertiary" stress.

415 $a LONG CONSONANTS $A Lisker gives examples of long consonants* but not an exhaustive list.

Krishnamurti (p.74ff ) says that in verbal bases all consonants occur long except the voiceless
fricatives, /h-voice/> /n-retrof lex/, and /yod/. It is not known whether these consonants occur
long in other contexts (i.e. at morpheme boundary, or due to external sandhi). They have not
been included in the segment inventory. tJHC)

415 $a LONG VOUELS. $A "In general, the difference between the short and long vowels is

important.... In the case of word-final vowels, however, the feature of length does not serve
as a distinctive characteristic of words. Instead, final vowels are highly variable in length,
generally short when the word is not final in a phrase, longer when it is in phrase-final
position. When the word-final vowel is at the end of a phrase, it is marked not only by
increased length, but also by... pitch." (p.xiv)

415 $a NASALIZED VOWELS $A Phrase-final sequences of vowel plus /n/ may optionally be realized as a

nasalized vowel. The only example given is /a-long.n/. (p.xviii) Moreover* the vowels in the
neighborhood of [w-nasalized] are "heavily nasalized." (p. 7)

415 $a PHONOLOGICAL WORD $A There are no vowel clusters. Otherwise different syllable types seem to

combine freely in words.

415 $a SYLLABLE $A (OVCKC) $A Lisker does not give a description of syllable structure nor any

information about phonotactics. The canonical formula given above is derived from the examples
he cites. The only examples cited that violate this canon are borrowings from Sanskrit and have
/p.r-flap/, /g.r-flap/, etc. [GHJ (This analysis is confirmed by Krishnamurti, p. 4.)

415 $a VOWEL HARMONY $A Lowering of vowels before /schwa, a-long/ in the following syllable can be

treated as a type of vowel lowering prosody. See note 66. tJHC) *

415 01 $A /t, d/ are "produced by bringing the tongue tip into contact with the upper front teeth or
with the roof of the mouth just behind the upper front teeth." (p. 7)

415 02 $A /t-retrof lex, d-retroflex/ are "produced by having the tongue tip make contact with the roof
of the mouth well back of the upper front teeth. Some speakers also curl the tongue tip up so
that a bit of the underside of the tongue makes contact with the roof of the mouth," (p. 8)

415 03 $A [r-f lap-retrof lex! is "like 'dd' of ladder,' often with tongue tip touching roof of mouth
well behind front teeth." (p.xvii)

415 06 $A /n/ "varies between 'n' of "ten 1 and 'n' of 'tenth. 1 " (p.xvii)

415 07 $A /r-flap/ is "like of 'cavity** but with tongue tip making contact with roof of mouth
just behind upper front teeth." (p.xviii)

415 09 $A Eiota] is "between 'i' of 'bit' and sound resembling both 1 i 1 of 'bit 1 and 'e' of 'bet;'
short," (p.xi i i )

415 10 - $A lepsilonl varies to higher-low.

415 11 $A lash-long] is "sometimes diphthongized beginning with very brief vowel like *a" of 'date 1
and shifting rapidly to vowel like 'a' of 'cat.'" Such optional diphthongization occurs "in
medial and final syllables of words." (p.xv)

415 12 $A [ash-long] varies to lower-mid. (p.xiv)

415 13 $A [ash-dot] is "between *u' of 'cut' and 'a' of 'cat; 1 short. 1 ' (p.xiii)

415 1l * $A [alpha-unrounded) "varies from 'u' of 'cut' to 'a' of 'father; 1 short." (p.xiii)

415 15 $A Ea-front-long] is "between 'a' of 'cat' and 'a' of 'father,' long." (p.xiii)

PAGE 234 STANFORD PHONOLOGY ARCHIVE

VOLUME 1 ~ SEGMENT INVENTORIES, GENERAL COMMENTS, FOOTNOTES Telugu 415

415 16 $A luJ is "between 'u' of 'put 1 and »oo» of 'poop; 1 short." (p.xiv)

415 17 $A ''The vowels /o, o-long, u, u-long/ are, at the beginning of words, usually pronounced with
an initial w-sound." (p.xv)

415 18 $A lupsilonl is "between 'u' of 1 put' and 'oa' of 'coat; 1 short." (p.xiv)

415 »» $A /beta-approximant/ is "like «v' of 'very 1 but lower lip is not drawn in." (p.xviil)

415 20 $A Krishnamurti says that the geminate /beta-approximant-long/ is labiodental tv] . (p. 5)

415 30 $A "Certain words, which in the main are Sanskrit in origin, are pronounced with aspirated stop
sounds in the careful speech of persons acquainted with the Sanskrit forms. In normal
colloquial style, however, the aspiration is usually not present, particularly if the stop
sound is not at the very beginning of a phrase." (p.xx)

415 3 ' $A Krishnamurti says that older */t-aspi rated/ has fallen together with /d-breathy voice/ in
modern Telugu. (p. 4)

415 32 $A "For the bulk of the Telugu vocabulary [tt/sl and tt/s-hacekll might be written with a

single symbol, for the occurrence of either Et/s-hacekl or Et/sJ is determined by the following
vowel. Before the front vowels .only E t/s-hacek] . . . is found, while before all other vowels '

there is only Et/s] In some dialects. . .only the sound It/s-hacekJ is used, and it is found

before all vowels. In representing the dialects with which we are concerned, however,

t/s-hacek) and Et/sJ are written differently because in some words (borrowed from Sanskrit)
[t/s-naceKV as well as It/sJ may be followed by the vowel tash-dotJ ." (p. 25)

415 33 $A /f/ is borrowed from English, (Krishnamurti, p. 4)

415

^^u 1 * 16 .^ 5 ^" 0 ^^ between tls3 and is-hacekll Is frequently blurred in the speech of many
Andhras, before the vowels /i , i-long, e, e-long/ many speakers will use either indifferently,
while before /ash-long/ they will use only ts-hacekl ." (p.xix)

415 35 $ A tn3 and Cn-retroflexl "are distinct in the speech of most educated speakers, but only In
intervocalic position," (p.xix)

415 36 $A In phrase final position /n/ may be dropped and the preceding vowel nasalized, (p.xviil)

415 37 $A "For many speakers the sound represented by /ash-long/ is ordinarily not distinguished from"
lepsi ion-long] . (p.xvi )

415 *° $A /d, d-retroflex, k, g/ are realized as (eth, r-f lap-retrof lex, k-lax, g-lax)
"intervocalically." (p.xvii)

415 61 $A The alveolar affricates are palatal before front vowels, (p.xix)

415 " $A M/ is realized as Ew-nasalizedJ in word final and intervocalic position. (Some loan words
do not undergo this change. Dialects outside of coastal Andhra Pradesh do not have the rule.

\ p.DT J J

415 63 $A /n/ is realized as tens] "before /k/ or /g/." (p.xvii)

415 $A /n/ is palatal before tt/s-hacek, d/z-hacek] . (Krishnamurti, p. 5)

415 65 $A /l/ is dental after /t/ and /d/. (p.xviii)

415 " $a vowels are lowered before /schwa, a-long/ in the following syllable. Intervenes /s-hacek,
t/s/ prevent lowering of /o, o-long/, and perhaps /e-long/. When final /schwa/ is elided in
connected speech before a vowel or glide in the next word, the lowered vowel remains. The
result is a marginal distinction between til and dotal , etc. which is quite important for
distinguishing forms in connected speech. (Lisker writes an apostrophe for elided /schwa/.)
Usker (p.xv, xvi, 3, 16, 17) and Kelley (who treats lowering as a prosodic feature) are n
complete agreement on the details of this process.

415 « $A /schwa/ is realized as lash-dot] after /t/s, s-hacek, d/z, yod/. (p.xiii)

415 «• $A /a-lons/ is realized as ta-f ront-lons) after palatal obstruents and /yod/.

4,5 69 ll^ 0 ^^^p!^r;! , " d as twl " follow1ns /schwa/ or /a - w and bef °- *— «•